TurkEmbed4Retrieval: Turkish Embedding Model for Retrieval Task

TL;DR

TurkEmbed4Retrieval is a specialized Turkish language embedding model optimized for retrieval tasks, achieving state-of-the-art performance by fine-tuning on the MS MARCO TR dataset with advanced training methods.

Contribution

The paper introduces TurkEmbed4Retrieval, a novel retrieval-focused Turkish embedding model that surpasses previous models like Turkish colBERT in retrieval benchmarks.

Findings

Outperforms Turkish colBERT by 19-26% on key metrics

Establishes new benchmark for Turkish retrieval tasks

Demonstrates effectiveness of advanced training techniques

Abstract

In this work, we introduce TurkEmbed4Retrieval, a retrieval specialized variant of the TurkEmbed model originally designed for Natural Language Inference (NLI) and Semantic Textual Similarity (STS) tasks. By fine-tuning the base model on the MS MARCO TR dataset using advanced training techniques, including Matryoshka representation learning and a tailored multiple negatives ranking loss, we achieve SOTA performance for Turkish retrieval tasks. Extensive experiments demonstrate that our model outperforms Turkish colBERT by 19,26% on key retrieval metrics for the Scifact TR dataset, thereby establishing a new benchmark for Turkish information retrieval.